At the edge of change: Native Phragmites persistence and nonnative arrival along the dam‐regulated Colorado River Authors: Emily C. Palmquist, Joel B. Sankey, Claudia DiMartini, Tayan Benson, Brooke Damon, Helen Fairley, Bradley J. Butterfield Published: 2026 ABSTRACT AbstractPremiseThe southwestern United States is at the forefront of nonnative Phragmites australis subsp. australis expansion. Native Phragmites australis subsp. americanus, likely in decline in North America, is ecologically and culturally important and of conservation interest. In the greater Grand Canyon region, the Glen Canyon Dam led to loss of habitat upstream and increases in Phragmites downstream. Recent increases correspond with changes to dam operations and the arrival of nonnative Phragmites. Determining the native status of Phragmites stands along the Colorado River and around Lake Powell can inform complex management choices and restoration planning.Methods Phragmites subspecies at 83 locations from southern Utah through northern Arizona, United States were genetically tested using real‐time PCR. Genetic diversity and structure were assessed using microsatellite loci for 13 native Phragmites stands. Morphological traits of southwestern native Phragmites were compared to those of nonnative Phragmites.ResultsNonnative Phragmites was identified at four of the 83 locations. Native Phragmites in the study area is genetically diverse with little regional genetic structure, but substantial genetic differentiation among sites. Southwestern native Phragmites is morphologically variable, but characteristics can be used together to distinguish it from nonnative Phragmites.ConclusionsThe greater Grand Canyon and Glen Canyon regions support a large, genetically diverse native Phragmites population, which is potentially threatened by nonnative Phragmites and changes to river management. Determining conditions that would lead to nonnative Phragmites invasion of desert wetlands will facilitate discussions about the risks and benefits of Phragmites management in these unique ecosystems. Phragmites australis (Cav.) Trin. ex Steud. (common reed; Poaceae) has been the focus of extensive scientific and management attention over the last three decades as nonnative stands spread across North America, changing the character of vast expanses of wetlands and riparian areas. This cosmopolitan species has native and nonnative lineages in North America, with the nonnative lineage expanding in the last 200 years and rapidly since the early 1900s (Saltonstall, 2002; Kettenring et al., 2012; Meyerson et al., 2025). The nonnative subspecies [Phragmites australis subsp. australis (Cav.) Trin. ex Steud.; hereafter, “nonnative Phragmites”] has spread from the east coast of North America to the west and south, replacing native stands, moving into ecosystems previously uninhabited by the species, and transforming wetlands throughout its invaded range (Saltonstall, 2002; Meyerson et al., 2025). The native subspecies (Phragmites australis subsp. americanus Saltonst., P.M. Peterson & Soreng; hereafter, “native Phragmites”) was widespread in North America before the expansion of nonnative Phragmites and is both ecologically and culturally important (Kiviat and Hamilton, 2001; Saltonstall, 2002; Kiviat, 2013; Meyerson et al., 2025). Research on all aspects of the native lineage is increasingly pressing as native stands are now uncommon or extirpated in many regions and are likely declining elsewhere (Meyerson et al., 2010a, 2025; Kulmatiski et al., 2011; Lambert et al., 2016). The Grand Canyon region in the southwestern United States (Figure 1) has long been home to native Phragmites and is at the leading edge of nonnative Phragmites spread (Lambert et al., 2016), making this a key area for understanding dynamics between the two subspecies. Map of Phragmites australis stands tested in this study (circles and triangles) and locations of P. australis subsp. australis and hybrids of P. australis subsp. australis and P. australis subsp. americanus identified by Lambert et al. (2016) occurring near the study area (squares). Numbered, light‐blue circles with a black outline (subsp. americanus*) indicate sites that were used to assess genetic structure and diversity; the numbers correspond to the site IDs in Table 1. The Colorado River flows from northeast to southwest; major tributaries are labeled. Map was generated using the U.S. Geological Survey (2025) shaded relief base map, Commission for Environmental Cooperation (2023) North American Atlas ‐ Lakes and Rivers layer, and the Esri (2025) Human Geography Map. U.S. National Park Service units included in sampling and Tribal lands along the Colorado River are shown. Native Phragmites has existed in the Grand Canyon region for more than 40,000 years (Hansen, 1978), enduring changing conditions, both natural and man‐made. It has been documented in preserved Pleistocene‐era dung at Rampart Cave (10,000 to >40,000 years ago), the Muav Caves (10,650 to 11,810 years ago), Stanton's Cave (>11,000 years ago), and Bechan Cave (11,670 to 13,505 years ago) (Hansen, 1978; Mead and Agenbroad, 1989), all within the greater Grand Canyon region. In the mid‐1900s, Phragmites was documented at seeps and springs in the Grand Canyon and in tributaries, but not along the edge of the Colorado River, the largest river flowing through the area (Clover and Jotter, 1944; Rand, 1957). In 1963, Glen Canyon Dam started operating, fundamentally changing flow patterns of the Colorado River by creating Lake Powell upstream and decreasing seasonal variability, increasing daily fluctuations, and eliminating large summer floods downstream that previously removed most vegetation (Topping et al., 2003; Palmquist et al., 2025b). Upstream of Glen Canyon Dam, Lake Powell flooded approximately 650 km2 of canyons containing seeps, springs, and perennial rivers that supported native Phragmites (Flowers, 1959; Woodbury et al., 1959) and replaced those habitats with a fluctuating reservoir shoreline largely unsuitable for Phragmites (Waring, 1992; Webb et al., 2007). In contrast, flow regulation downstream of the dam starting in 1963 led to a general increase in riparian vegetation along the Colorado River in the Grand Canyon, including both native and nonnative plant species (Turner and Karpiscak, 1980; Sankey et al., 2015; Durning et al., 2021). Phragmites became a common part of the regulated Colorado River riparian community, and persisted in seeps, springs, and tributaries (Phillips and Phillips, 1979; Turner and Karpiscak, 1980; Stevens et al., 1995). Since 2002, Phragmites has expanded farther along the dam‐regulated Colorado River (Durning et al., 2021). At this same time, nonnative Phragmites arrived and became dominant in northern Utah (approximately 600 km north of the Grand Canyon region) (Kulmatiski et al., 2011). An assessment of the distribution of native and nonnative Phragmites from 2006 to 2015 found new records of nonnative Phragmites in the Colorado River basin to the south and west of the Grand Canyon region (Figure 1; Lambert et al., 2016), making the spread of nonnative Phragmites upstream into the Grand Canyon likely. Additionally, large stands of a hybrid of native and nonnative Phragmites were discovered in the Las Vegas Wash drainage area, Nevada, a wetland area to the west of the Grand Canyon region (purple square in Figure 1; Saltonstall et al., 2016). As such, the recent expansion of Phragmites along the Colorado River in the Grand Canyon could be related to ongoing river regulation, the arrival of nonnative Phragmites, or a combination of both factors. Shifting distributions of native, nonnative, and hybrid Phragmites in western North America have led to ongoing management concerns and attempts to remove large stands (Kulmatiski et al., 2011; Saltonstall et al., 2016; Long et al., 2017b; Abella et al., 2024). Nonnative Phragmites can substantially degrade ecosystems in fundamental ways, including nutrient cycling, sedimentation, species diversity, and vegetation structural complexity (Meyerson et al., 2025). Allowing nonnative Phragmites to expand in rare, desert wetland and riparian habitats could lead to socially unacceptable declines in biodiversity, increased fire hazard, and general ecosystem degradation (Abella et al., 2024; Meyerson et al., 2025). There is also the possibility of losing native Phragmites entirely either through competition or genetic swamping (Saltonstall et al., 2016; Meyerson et al., 2025), thus losing its structural and ecological traits. In the greater Grand Canyon region, the occurrence of nonnative Phragmites could prompt land management agencies to remove the nonnative plants and restore with native Phragmites. If dam operations are supporting large stands of native Phragmites, changes to those operations could alter this unique population and/or concurrently provide suitable habitat for nonnative expansion (Butterfield and Palmquist, 2024; Palmquist et al., 2025c). Detailed evaluations of the distribution and genetic make‐up of native and nonnative Phragmites in the Grand Canyon region can assist with complex management considerations. Management of Phragmites in the Grand Canyon region has direct implications for cultural connections to Phragmites. The Grand Canyon region is a sacred landscape (Hirst, 2016) to several southwestern Native American Tribes, and many southwestern Indigenous peoples (more than 45 Tribes) consider Phragmites an important spiritual, ecological, and utilitarian species (Elmore, 1944; Whiting, 1985; Mayes and Lacy, 1989; Moerman, 1998; Kiviat and Hamilton, 2001; Anderson, 2005). Some of the most commonly documented uses are as components of arrow shafts, smoking accessories, flutes or whistles, woven mats, and fire‐making equipment (Elmore, 1944; Whiting, 1985; Kiviat and Hamilton, 2001). Parts of the plant are edible, but sugar derived from the plant itself and the exudate from associated aphids are the most commonly mentioned foods, reflecting its historical importance in the sugar‐limited ecosystems of the Southwest (Whiting, 1985; Kiviat and Hamilton, 2001; Anderson, 2005). Native Phragmites is used in ceremonies (Elmore, 1944; Moerman, 1998) and appears in oral traditions (Vecsey, 1983; Bruchac, 1991). For example, the reed served as a ladder that enabled people to move from one world to the next in the emergence stories of multiple cultures (Vecsey, 1983; Bruchac, 1991). Native Phragmites, then, plays roles in many Indigenous cultures as tangible products, ceremonial components, and as a key figure in creation stories. When making management decisions about a culturally and ecologically valuable plant in a sacred landscape, confidence in subspecies identification before implementing management actions supports better communication, decision‐making, and collaboration (Hoelting et al., 2025). Genetic testing remains the most reliable method for distinguishing among Phragmites subspecies (Meyerson et al., 2025), and documenting existing genetic structure and diversity of native Phragmites can be used in restoration planning. Thus, a genetic assessment of Phragmites in the greater Grand Canyon region could increase certainty in subspecies identification and generate data that can be used for developing restoration materials. However, reliable field identification (without genetic testing) in the Grand Canyon region could facilitate rapid management responses. Existing keys are derived from the extensive literature on North American Phragmites, so represent the breadth of nonnative Phragmites morphology well (Allred and Saltonstall, 2007; Swearington and Saltonstall, 2012; McTavish et al., 2023). They do not, however, equally represent the morphological diversity of southwestern Phragmites, simply because fewer studies have focused on southwestern native Phragmites (Meyerson et al., 2025). Thus, a paired genetic and morphological assessment of southwestern native Phragmites could help determine whether morphological traits used to distinguish the subspecies elsewhere can also be reliably used in the southwestern United States. This study aimed to (1) evaluate the distribution of native and nonnative Phragmites in the Grand Canyon region, (2) determine the genetic diversity and structure of native Phragmites for restoration planning, and (3) assess the utility of morphological characters to distinguish between native and nonnative Phragmites in southwestern landscapes. The results of this study aim to provide management‐relevant information that can be used to facilitate conversations among land managers and stakeholders on ways to protect native Phragmites and address the spread of nonnative Phragmites. MATERIALS AND METHODS Study area The study area consists of the Colorado River and its tributaries from Arches National Park in southern Utah to the western end of Grand Canyon National Park, Arizona, predominately upstream of Lake Mead (Figure 1). This part of the Colorado River basin is characterized by networks of large, deep canyons that both connect and divide ecological communities (Stevens, 2012; Palmquist et al., 2021). In this region, the Colorado River flows through desert landscapes and is the largest source of water. Other wetland habitats include smaller perennial and intermittent rivers and creeks and isolated springs separated by dryland, upland vegetation (Stevens and Meretsky, 2008; Stortz et al., 2018). Water courses run along the canyon bottoms, while steep cliffs hinder cross‐landscape movement. The Colorado River flows from northeast to southwest with many small to medium‐sized tributaries flowing to the Colorado River from higher elevation plateaus (Stortz et al., 2018). Glen Canyon Dam, approximately 25 km upstream of the Paria River confluence, forms Lake Powell. Hoover Dam, approximately 575 km downstream of Glen Canyon Dam, forms Lake Mead. In between, Colorado River flows are controlled by Glen Canyon Dam, with flows typically ranging from 226 to 707 m3/s (Topping et al., 2003), except during occasional experimental releases of higher or lower flow from the dam (U.S. Department of Interior, 2016). Lake Powell inundated a network of canyon tributaries to the Colorado River and, at full pool, has a surface area of 644 km2 and is 299 km long (Deemer et al., 2025). The shoreline of Lake Powell fluctuates with inflows and dam releases and largely meets sandstone bedrock (Wildman et al., 2011). Lake Powell has been declining over the last two decades and has not had a surface area larger than about 470 km2 since 2016 and no more than 350 km2 since 2021 (U.S. Bureau of Reclamation, 2026). Upstream of Glen Canyon Dam, Phragmites occurs primarily in tributaries to the Colorado River (e.g., San Juan River, Escalante River) near confluences of those tributaries and the Colorado River, and in isolated canyon seeps and springs (Flowers, 1959; Woodbury et al., 1959; Spence, 1992; Tendick et al., 2017). Pre‐dam (pre‐reservoir), Phragmites was not a dominant species, but was frequent along streams in Glen Canyon (Flowers, 1959) and was estimated to cover approximately 8.8 ha along the Colorado River within the footprint of Lake Powell (Woodbury et al., 1959). The creation of Lake Powell reduced the amount of available habitat for Phragmites because the shoreline of the lake is largely composed of sandstone bedrock and therefore not suitable habitat (Waring, 1992; Webb et al., 2007). Post‐dam, Phragmites persists in areas not inundated by Lake Powell (Spence, 1992; Tendick et al., 2017), but has largely not colonized the fluctuating shoreline except in tributary inflows (Waring, 1992). The low level of Lake Powell over the last two decades has exposed deltaic sediment deposits in the reservoir backwater and inflow areas with fine sediment and nutrients that could be suitable habitat for nonnative Phragmites (Deemer et al., 2025; Kasprak et al., 2026). Recent vegetation maps of Glen Canyon National Recreation Area (National Recreation Area) indicate that Phragmites is a component of four vegetation classes that, in total, cover approximately 3700 ha, although Phragmites cover is much lower because it is not a dominant species in these communities (Tendick et al., 2017). Between Lake Powell and Lake Mead, Phragmites occurs along the Colorado River and in tributary springs (Phillips and Phillips, 1979; Durning et al., 2018; Palmquist et al., 2023). Populations of Phragmites in tributaries were already long‐established before the Glen Canyon Dam (Rand, 1957), but Phragmites colonized the Colorado River after dam operations began in 1963 (Stevens et al., 1995). Early periods of large daily flow fluctuations, as much as 790 m3/s, led to the development of fluvial marshes, including those characterized by Phragmites (Stevens et al., 1995). Unusually large flooding and high flows from 1983 through 1986 removed most of these marshes, but they recolonized quickly (Stevens and Waring, 1985; Stevens et al., 1995). By 1991, marshes that included stands of Phragmites covered 9.0 ha (Stevens et al., 1995). Changes to dam operations starting in 1991 reduced daily flow fluctuations to no more than 226 m3/s per day, which has corresponded with a decline in marshes (Kearsley and Ayers, 1996). As of 2013, Phragmites covered approximately 3.7% (4.40 ha) of the vegetated area along the Colorado River between Glen Canyon Dam and Lake Mead (Durning et al., 2021) and is still one of the dominant species on large eddy sandbars (Palmquist et al., 2023). Native and nonnative Phragmites can spread through seed, rhizomes, and rooting from stem nodes (Meyerson et al., 2025). In the western United States, nonnative Phragmites seems to spread more by seed than vegetatively, while spread of the native Phragmites is primarily vegetative (Kettenring and Mock, 2012). Phragmites can move through a watershed along river channels via water and wind dispersal of seed and from sprouting from stem and rhizome fragments. It can move across watersheds via seeds transported by wind, vehicles, or animals. In the eastern United States, nonnative Phragmites expansion has been associated with the drainage ditches, disturbance, and higher nutrient loads of roads and canals (Jodoin et al., 2008; Kirk et al., 2011). The climate of the study region is warm and dry, with higher temperatures at lower elevations (Fick and Hijmans, 2017). The elevations of the sampled sites range from 370 to 1280 m a.s.l. (Table 1). At the farthest north and upstream sample site, mean annual temperature (MAT) is lowest and mean annual precipitation (MAP) greatest (12.6°C, 226 mm). In the Grand Canyon, near the confluence of the Colorado River and the Little Colorado River, MAT is 17.1°C, and MAP is 158 mm (Fick and Hijmans, 2017). At the farthest downstream sample site, MAT is 23.0°C, and MAP is 148 mm (Fick and Hijmans, 2017). Table 1Collection site information, sample sizes, and real‐time qPCR threshold cycle (Ct) values. Site ID = code used for reference in main text. Site code = short, descriptive code for site. Location, elevation (a.s.l.) = full description of site. Locations are based on river miles from Lees Ferry and the river side when facing downstream and used by convention in all maps. For example, river mile 2.1L indicates the site is 2.1 miles downstream of Lees Ferry and on the left side of the river. Negative values indicate distance upstream of Lees Ferry. Location coordinates x = x coordinate of location in decimal degrees (WGS, 1984), y = y coordinate of location in decimal degrees (WGS, 1984). Collector = collector name and if available, specimen number. Herbarium = herbarium housing the voucher collection. Herbarium codes: GLCA = Glen Canyon National Recreation Area herbarium, GCNP = Grand Canyon National Park herbarium, DES = Desert Botanical Garden herbarium, Hualapai = Hualapai Indian Tribe herbarium. N qPCR (SSR) = sample sizes for qPCR and simple sequence repeat (microsatellite, SSR) analysis. Ct mean (SD) = Ct means and standard deviations for the AMAU and AMER probes (NA = no amplification).Site IDSite codeLocation, elevation (a.s.l.)Location coordinatesCollectorHerbariumN qPCR (SSR)Ct mean (SD)xyAMAUAMER Phragmites australis subsp. australis Type MType M SalinasPositive controlA. Lambert–NA20.2 (1.5)NA (NA)1NN CourthouseUSA. Utah. Grand County. Arches National Park. Courthouse Wash, 1280 m–109.6437275138.6727424K. Quinn–1 (0)18.5 (0.4)NA (NA)2Arches 1 AUSA. Utah. Grand County. Arches National Park. Above Sevenmile in Courthouse Wash, upper population, 1280 m.–109.6437275138.6727424C.J. & J.L.–1 (1)20.4 (0.5)NA (NA)3Arches 2 AUSA. Utah. Grand County. Arches National Park. Sevenmile Courthouse Wash, middle population, 1275 m.–109.6409889138.66930055C.J. & J.L.–1 (1)19.7 (0.6)NA (NA)4Paria Riffle 0.3 RUSA. Arizona. Coconino County. Glen Canyon National Recreation Area. Along Colorado River at river mile 0.3 R, Paria Riffle, 950 m.–111.5922150636.8638561E. Palmquist 315GLCA1 (1)18.8 (0.4)NA (NA)5Paria BeachUSA. Arizona. Coconino County. Grand Canyon National Park. Along Colorado River at river mile 1.2 R, Paria Beach, 950 m.–111.5922150636.8638561E. Palmquist 304GCNP2 (1)18.5 (0.2), 19.6 (0.3)NA (NA) P. a. subsp. australis X P. a. subsp. americanus HybridHybrid Las Vegas WashPositive ControlA. Lambert–0 (1)–– Phragmites australis subsp. americanus Type BType B CalientePositive ControlA. Lambert–1 (1)21.2 (1.4)24.7 (2.8)Type HType H North Shore BridgePositive ControlA. Lambert–0 (1)––6Arches 1BUSA. Utah. Grand County. Arches National Park. Above Sevenmile in Courthouse Wash, upper population, 1280 m.–109.6437275138.6727424C.J. & J.L.–1 (0)18.6 (0.3)20.5 (0.6)7Arches 2BUSA. Utah. Grand County. Arches National Park. Sevenmile Courthouse Wash middle population, 1275 m.–109.6409889138.66930055C.J. & J.L.–1 (0)20.5 (0)24.5 (0.4)8Dirty Devil Scenic Rest AreaUSA. Utah. Garfield County. Glen Canyon National Recreation Area. In the wash south of Dirty Devil Scenic Rest Area, 1110 m.–110.3985353737.90045253G. WattsGLCA1 (1)20.5 (0)26.4 (2.9)9North WashUSA. Utah. Garfield County. Glen Canyon National Recreation Area. North Wash, on the southwest side of Highway 95 about 400 m south of the Glen Canyon National Recreation Area boundary, 1140 m.–110.4527479737.9044738G. WattsGLCA1 (1)21.3 (0.5)26.2 (0.9)10Hobie Cat Road BullfrogUSA. Utah. Kane County. Glen Canyon National Recreation Area. Southwest of Bullfrog Campground, in wash next to road that goes to Hobie Cat Beach, 1120 m.–110.7250353237.51485611G. WattsGLCA1 (1)21.3 (0.5)26.2 (0.9)11Bullfrog Boat Rental ParkingUSA. Utah. Kane County. Glen Canyon National Recreation Area. Near Bullfrog Marina, next to North Marina Road boat rental parking area, 1100 m.–110.7396335137.51901971G. WattsGLCA1 (1)21.5 (0.1)23.3 (0)12Escalante‐Neon CanyonUSA. Utah. Garfield County. Glen Canyon National Recreation Area. Escalante River near the confluence with Neon Canyon, 1380 m.–111.1682752737.60651707J. Hornfeldt–1 (15)20.9 (0)25.8 (0.1)13San Juan RM 52USA. Utah. San Juan County. Glen Canyon National Recreation Area. San Juan River at river mile 52 R, 1180 m.–110.0095276637.22053372J. HornfeldtGLCA1 (15)19.4 (0.4)21.1 (2.4)14San Juan SlickhornUSA. Utah. San Juan County. Glen Canyon National Recreation Area. San Juan River near river mile 66 R, Slickhorn D campsite, upstream of Slickhorn Rapids and at the confluence with Slickhorn Canyon, 1150 m.–110.1482337437.31657488J. Hornfeldt–1 (1)25.7 (0.4)28.8 (1.4)15Escalante ConfluenceUSA. Utah. San Juan County. Glen Canyon National Recreation Area. Lake Powell near the confluence with Escalante River, on north side of lake, 1130 m.–110.8655752937.29441542T. PrestonGLCA1 (1)22.2 (0.2)23.1 (0.4)16Lake R Bouy 67USA. Utah. Kane County. Glen Canyon National Recreation Area. Small side canyon off Lake Powell near buoy 67 on west side of lake, opposite Ribbon Canyon, 1125 m.–110.8687750137.26422131T. PrestonGLCA1 (1)23.5 (0.1)25.5 (0.2)17Lake L Bouy 59USA. Utah. Kane County. Glen Canyon National Recreation Area. Small side canyon off Lake Powell near buoy 59 on east side of lake just upstream of the confluence with San Juan River, 1125 m.–110.8844181937.19712575T. PrestonGLCA1 (1)23.5 (0.1)25.8 (0.4)18San Juan ConfluenceUSA. Utah. San Juan County. Glen Canyon National Recreation Area. Small side canyon on north side of the confluence of the Colorado and San Juan Rivers in Lake Powell, 1120 m.–110.8881552537.18610825T. PrestonGLCA1 (1)21.5 (0.3)22.3 (0.5)19Reflection CanyonUSA. Utah. Kane County. Glen Canyon National Recreation Area. Reflection Canyon, 4 km from entering the canyon from Lake Powell on north side of canyon, 1115 m.–110.9148956337.19093659T. PrestonGLCA1 (1)21.7 (0.3)22.8 (0.3)20Warm Creek Bay NE CanyonUSA. Utah. Kane County. Glen Canyon National Recreation Area. Warm Creek Bay of Lake Powell in canyon on east side of bay, 1100 m.–111.4396661437.06813507G. WattsGLCA1 (1)23.2 (0.1)25.2 (1.2)21Warm Creek Bay SE CanyonUSA. Utah. Kane County. Glen Canyon National Recreation Area. Warm Creek Bay of Lake Powell in canyon on southeast end of bay, 1110 m.–111.4277747537.05777194G. WattsGLCA1 (15)24.5 (0.8)26.2 (0.3)22Canyon NE of Skylight ArchUSA. Utah. Kane County. Glen Canyon National Recreation Area. In canyon on south side of Wahweap Bay of Lake Powell, northeast of Skylight Arch, 1140 m.–111.5681344237.02605474G. WattsGLCA1 (1)21.1 (0.3)23 (0.2)23−13.7 LUSA. Arizona. Coconino County. Glen Canyon National Recreation Area. Along Colorado River at river mile –13.7 L, 950 m.–111.5000902336.9156177G. WattsGLCA1 (1)22.8 (0.1)25.6 (0.4)24Honey Draw −13.2 RUSA. Arizona. Coconino County. Glen Canyon National Recreation Area. Along Colorado River across from Honey Draw at river mile –13.2 R, 950 m.–111.5011412636.90966811G. WattsGLCA1 (1)23.4 (0.1)26.4 (2.4)25−11.8 RUSA. Arizona. Coconino County. Glen Canyon National Recreation Area. Along Colorado River at river mile –11.8 R, 950 m.–111.5165352936.90201287G. WattsGLCA1 (1)24.4 (0.2)26.4 (0.1)26−10.6 RUSA. Arizona. Coconino County. Glen Canyon National Recreation Area. Along Colorado River at river mile –10.6 R, 950 m.–111.5264572636.89196448G. WattsGLCA1 (1)23.2 (0.2)25.4 (0.7)27−9.9 RUSA. Arizona. Coconino County. Glen Canyon National Recreation Area. Along Colorado River near river mile –9.9 R, 940 m.–111.5218863536.88451878G. WattsGLCA1 (1)23.5 (0.3)26.2 (1.1)28 Leopard Frog Marsh −9.0 R USA. Arizona. Coconino County. Glen Canyon National Recreation Area. Along Colorado River at river mile –9.0 R, 940 m.–111.5163631236.87719465G. WattsGLCA1 (15)18.8 (0.3)21.4 (1.5)29−8.3 RUSA. Arizona. Coconino County. Glen Canyon National Recreation Area. Along Colorado River at river mile –8.3 R, 940 m.–111.5257173936.87868248G. WattsGLCA1 (1)20.2 (0.4)27.9 (0.6)30−7.8 LUSA. Arizona. Coconino County. Glen Canyon National Recreation Area. Along Colorado River at river mile –7.8 L, 940 m.–111.5339052336.87925288G. WattsGLCA1 (1)19.4 (0.1)22.9 (3.9)31 Lunch Beach −7.1 L USA. Arizona. Coconino County. Glen Canyon National Recreation Area. Along Colorado River at Lunch Beach, river mile –7.1 L, 940 m.–111.5492691836.87428091E. Palmquist 284GLCA1 (1)20.5 (0.3)23.4 (1.8)32 Hidden Slough −6.6 R USA. Arizona. Coconino County. Glen Canyon National Recreation Area. Along Colorado River at river mile –6.6 R, 940 m.–111.5551155936.87367719E. Palmquist 285GLCA1 (15)20 (1.5)22.9 (2.9)33−5.5 RUSA. Arizona. Coconino County. Glen Canyon National Recreation Area. Along Colorado River at river mile –5.5 R, 940 m.–111.56201336.86423659G. WattsGLCA1 (1)19.8 (0.4)25.6 (2.4)34 Lees Ferry −0.1 R USA. Arizona. Coconino County. Glen Canyon National Recreation Area. Colorado River on west side of Lees Ferry boat ramp, river mile –0.1 R, 940 m.–111.5876231136.86550066G.WattsGLCA1 (1)19.4 (1.4)23.9 (2.2)35Paria River CurveUSA. Arizona. Coconino County. Glen Canyon National Recreation Area. Paria River near curve in river about 1 km downstream from Glen Canyon National Recreation Area boundary, 950 m.–111.6055161136.88379668C. DiMartini & T. BensonGLCA1 (1)19.3 (1.1)26.9 (1.8)36Paria River GageUSA. Arizona. Coconino County. Glen Canyon National Recreation Area. Paria River about 150 m upstream of USGS water monitoring gage: Paria River at Lees Ferry, AZ 09382000, 950 m.–111.5960873936.87316821C. DiMartini & T. BensonGLCA1 (1)18 (0.2)21 (1.7)37Paria River LonelyUSA. Arizona. Coconino County. Glen Canyon National Recreation Area. Paria River immediately south of Lonely Dell historic site, 960 m.–111.5944535436.86677455C. DiMartini & T. BensonGLCA1 (1)17.8 (0.3)19.5 (1.8)382.1 LUSA. Arizona. Coconino County. Grand Canyon National Park. Along Colorado River at river mile 2.1 L, 960 m.–111.6158909736.84745586E. Palmquist 305GCNP1 (1)19.7 (0.4)25.1 (2.4)3910.7 LUSA. Arizona. Coconino County. Grand Canyon National Park. Along Colorado River at river mile 10.7 L, 940 m.–111.6857364936.7455822E. Palmquist 311GCNP1 (1)19.3 (0.5)24.8 (1.7)40Hot Na Na 16.6 LUSA. Arizona. Coconino County. Grand Canyon National Park. Along Colorado River at river mile 16.6 L, Hot Na Na Camp, 920 m.–111.7391110836.68109512E. Palmquist 312GCNP1 (1)17.5 (0.2)19 (0.5)41Vasey's Paradise 32.2 RUSA. Arizona. Coconino County. Grand Canyon National Park. Along Colorado River at river mile 32.2 R, Vasey's Paradise, 890 m.–111.8574551536.49805784E. Palmquist 293GCNP1 (15)20.4 (0.3)20.6 (1.7)42Lower Buck Farm 41.4 RUSA. Arizona. Coconino County. Grand Canyon National Park. Along Colorado River at river mile 41.4 R, Lower Buck Farm, 860 m.–111.8827570336.40170417E. Palmquist 294GCNP1 (1)17.1 (0.1)18.4 (0.2)4351.2 LUSA. Arizona. Coconino County. Grand Canyon National Park. Along Colorado River at river mile 51.2 L, 860 m.–111.8597376536.32318355E. Palmquist 286GCNP2 (1)17.6 (0.1), 18.7 (0.3)18.3 (0.4), 20.8 (1.6)4453.3 RUSA. Arizona. Coconino County. Grand Canyon National Park. Along Colorado River at river mile 53.3 R, 850 m.–111.8591354736.29812246E. Palmquist & D. Sanderson–1 (1)18.7 (0.4)21.6 (1.4)45NankoweapUSA. Arizona. Coconino County. Grand Canyon National Park. Nankoweap Creek in Nankoweap Canyon near where the Nankoweap Trail descends to canyon bottom, about 2.5 miles up creek from the confluence with Colorado River, 1060 m.–111.8899876536.2801972E. Palmquist 316GCNP1 (15)18.1 (0.1)20.2 (1.9)46Kwagunt Marsh 55.9 RUSA. Arizona. Coconino County. Grand Canyon National Park. Along Colorado River at river mile 55.9 R, Kwagunt Marsh, 850 m.–111.834276136.26898856E. Palmquist 295GCNP1 (1)19.1 (0.7)19.9 (2)4756.7 RUSA. Arizona. Coconino County. Grand Canyon National Park. Along Colorado River at river mile 56.7 R, 840 m.–111.8259061336.26021634E. Palmquist 306GCNP1 (1)20.5 (0.3), 18.9 (0.3)25 (1.7), 23 (0.9)4861.5 LUSA. Arizona. Coconino County. Grand Canyon National Park. Along Colorado River at river mile 61.5 L, 830 m.–111.7990166836.19575429C. DiMartini & A. LippmanGCNP1 (1)17.9 (0.2)19.7 (0.8)49Across from the LCR 62.0 RUSA. Arizona. Coconino County. Grand Canyon National Park, north side of island in Colorado River across from confluence of Little Colorado River, 830 m.–111.8042816736.19182025W. Hodgson & C. CannonDES1 (15)18.8 (0.4)20.5 (0.4)5062.6 RUSA. Arizona. Coconino County. Grand Canyon National Park. Along Colorado River at river mile 62.6 R, 820 m.–111.8115471336.18540781E. Palmquist, R. Gurule, D. Sanderson–1 (1)19.3 (0.5)20.5 (1.4)5168.4 RUSA. Arizona. Coconino County. Grand Canyon National Park. Along Colorado River at river mile 68.4 R, downstream end of Upper Tanner Camp, 820 m.–111.8268802236.10698459E. Palmquist 287GCNP1 (1)18.1 (0.2)20.5 (1.7)5268.5 LUSA. Arizona. Coconino County. Grand Canyon National Park. Along Colorado River at river mile 68.5 L, 820 m.–111.8265157336.10406444E. Palmquist 307GCNP1 (1)19.5 (0.2)22 (0.9)5369.0 LUSA. Arizona. Coconino County. Grand Canyon National Park. Along Colorado River at river mile 69.0 L, on debris fan at mouth of Tanner Canyon, 820 m.–111.8341412536.10318245E. Palmquist 308GCNP2 (1)18.3 (0.9), 19.1 (0.6)21.5 (2), 24.1 (2.5)5471.4 LUSA. Arizona. Coconino County. Grand Canyon National Park. Along Colorado River at river mile 71.4 L, upstream of Cardenas Camp, 810 m.–111.8610264636.08625765E. Palmquist 296GCNP1 (1)20.4 (0.2)25.8 (0.9)5572.2 LUSA. Arizona. Coconino County. Grand Canyon National Park. Along Colorado River at river mile 72.2 L, 800 m.–111.8751388736.08597972D. Sanderson, E. Palmquist, R. Gurule–2 (2)18.5 (0.2), 19.7 (0.2)20.6 (0.2), 22.4 (3.2)5672.5 LUSA. Arizona. Coconino County. Grand Canyon National Park. Along Colorado River at river mile 72.5 L, 800 m.–111.877298236.08333121E. Palmquist 303GCNP2 (1)17.9 (0.2), 19.2 (0.1)19.9 (0.8), 22.2 (1.3)5776.4 LUSA. Arizona. Coconino County. Grand Canyon National Park. Along Colorado River at river mile 76.4 L, on debris fan formed by Papago Creek, 790 m.–111.9065578736.04551243E. Palmquist 310GCNP1 (1)17.7 (0.2)19.8 (1.1)58Bright Angel CreekUSA. Arizona. Coconino County. Grand Canyon National Park. Bright Angel Creek near the trail turn off for Ribbon Falls from North Kaibab Trail, 1130 m.–112.0521667336.15939634E. Palmquist 317GCNP1 (15)18.7 (0.2)20.9 (2.4)59Havasupai GardensUSA. Arizona. Coconino County. Grand Canyon National Park. Along Bright Angel Trail at Havasupai Gardens, 1150 m.–112.127325336.07834133C. DiMartini & T. BensonGCNP1 (15)19.4 (2.3)20.9 (1.4)60Granite Camp 93.8 LUSA. Arizona. Coconino County. Grand Canyon National Park. Along Colorado River at river mile 93.8 L, Granite Camp, 720 m.–112.1818168236.09747481E. Palmquist 297GCNP1 (14)18.5 (0.1)20.2 (0.1)61Above Crystal 98.6 RUSA. Arizona. Coconino County. Grand Canyon National Park. Along Colorado River at river mile 98.6 R, just upstream of Crystal Camp, 720 m.–112.2405310636.13428163E. Palmquist 309GCNP2 (1)18.1 (0.1), 18.2 (0.2)19.9 (0.4), 20.1 (0.2)62120.4 RUSA. Arizona. Coconino County. Grand Canyon National Park. Along Colorado River at river mile 120.4 R, 650 m.–112.4687297936.23961409A. Lippman, H. Dutton, E. Palmquist–1 (1)19.2 (0.3)23.6 (0.2)63Forster 123.3 LUSA. Arizona. Coconino County. Grand Canyon National Park. Along Colorado River at river mile 123.3 L, Upper Forster Camp, 630 m.–112.5170382336.24453823E. Palmquist 298GCNP1 (1)18.1 (0.4)19.3 (1.1)64129.0 RUSA. Arizona. Coconino County. Grand Canyon National Park. Along Colorado River at river mile 129.0 R, 620 m.–112.4873532836.30831444E. Palmquist–1 (1)19.7 (0.2)22.2 (1.5)65142.7 RUSA. Arizona. Coconino County. Grand Canyon National Park. Along Colorado River at river mile 142.7 R, 580 m.–112.6073530836.39481984E. Palmquist 313GCNP1 (1)18.2 (0.3)19.7 (1.5)66Kanab 144.0 RUSA. Arizona. Coconino County. Grand Canyon National Park. Along Colorado River at river mile 144.0 R, mouth of Kanab Creek, 580 m.–112.6299955936.39276893E. Palmquist 314GCNP1 (1)17.8 (0.3)21.5 (3.5)67167.7 LUSA. Arizona. Coconino County. Grand Canyon National Park/Hualapai Indian Reservation. Along Colorado River at river mile 167.7 L, 530 m.–112.9058410136.25427789D. Sanderson, C. Keediniihii, C. DiMartiniHualapai1 (1)18.4 (0.4)20.5 (2)68Stairway 171.6 RUSA. Arizona. Mohave County. Grand Canyon National Park. Along Colorado River at river mile 171.6 R, Stairway Canyon Camp, 530 m.–112.9601041536.23669464E. Palmquist 288GCNP1 (15)23 (0.1)28.3 (1.4)69172 Mile 172.6 LUSA. Arizona. Coconino County. Grand Canyon National Park/Hualapai Indian Reservation. Along Colorado River at river mile 172.6 L, 530 m.–112.977122436.23229804E. Palmquist 299Hualapai1 (1)20 (0.4)22.8 (1.3)70186.8 LUSA. Arizona. Coconino County. Grand Canyon National Park/Hualapai Indian Reservation. Along Colorado River at river mile 186.8 L, 490 m.–113.18123736.15782086E. Palmquist 289Hualapai2 (1)20 (0.3), 21.9 (0.4)22.9 (2.2), 29.6 (1.6)71190.7 LUSA. Arizona. Coconino County. Grand Canyon National Park/Hualapai Indian Reservation. Along Colorado River at river mile 190.7 L, 480 m.–113.2062205136.11418327E. Palmquist–1 (1)18.4 (0.1)21 (0.8)72191.3 RUSA. Arizona. Mohave County. Grand Canyon National Park. Along Colorado River at river mile 191.3 R, 480 m.–113.214178536.11003793E. Palmquist 318GCNP1 (1)17.5 (0.2)18.3 (0.6)73194.6 LUSA. Arizona. Coconino County. Grand Canyon National Park/Hualapai Indian Reservation. Along Colorado River at river mile 194.6 L, Hualapai Acres Camp, 480 m.–113.2531680636.08898154E. Palmquist 300Hualapai1 (1)19.1 (0.1)23.6 (2.3)74195.5 RUSA. Arizona. Mohave County. Grand Canyon National Park. Along Colorado River at river mile 195.5 R, 470 m.–113.2697071436.09343582E. Palmquist 290GCNP1 (1)20.6 (2.2)21.7 (0.7)75198.3 RUSA. Arizona. Mohave County. Grand Canyon National Park. Along Colorado River at river mile 198.3 R, 470 m.–113.3162144136.09917227E. Palmquist–1 (1)18.3 (0)23.6 (3.1)76201 LUSA. Arizona. Coconino County. Grand Canyon National Park/Hualapai Indian Reservation. Along Colorado River at river mile 201.0 L, 460 m.–113.3335754236.06624698D. Sanderson, A. Lippman, C. DiMartini–1 (1)19.1 (0.1)21.4 (0.5)77207.2 RUSA. Arizona. Mohave County. Grand Canyon National Park. Along Colorado River at river mile 207.2 R, 450 m.–113.3397291535.98667242E. Palmquist 291GCNP2 (1)21 (0.5), 21.5 (0.2)27.4 (1.4), 29.3 (1.2)78211.8 RUSA. Arizona. Mohave County. Grand Canyon National Park. Along Colorado River at river mile 211.8 R, 440 m.–113.3260982835.93442938D. Sanderson, A. Lippman, E. Palmquist–2 (1)17.2 (0.4), 19.8 (1)18.9 (0.1), 19.7 (2.7)79214.8 RUSA. Arizona. Mohave County. Grand Canyon National Park. Along Colorado River at river mile 214.8 R, 430 m.–113.3220958535.89759904D. Sanderson, A. Lippman, E. Palmquist–1 (1)17.8 (0.1)23.1 (2.4)80Trail Canyon 219.6 RUSA. Arizona. Mohave County. Grand Canyon National Park. Along Colorado River at river mile 219.6 R, Trail Canyon camp, 420 m.–113.3276262135.83823929E. Palmquist–1 (1)20.4 (0.2)24 (2.1)81Middle Gorilla 220.1 RUSA. Arizona. Mohave County. Grand Canyon National Park. Along Colorado River at river mile 220.1 R, Middle Gorilla Camp, 420 m.–113.3264041635.83203166E. Palmquist 301GCNP1 (1)21.5 (0.3)29.2 (0.4)82225 Mile Camp 225.5 RUSA. Arizona. Mohave County. Grand Canyon National Park. Along Colorado River at river mile 225.5 R, 225 Mile Camp, 410 m.–113.3696774335.77167549E. Palmquist 302GCNP2 (1)20.3 (0.7), 19.3 (0.7)26.3 (1.8), 22.6 (2.2)83243.3 RUSA. Arizona. Mohave County. Grand Canyon National Park. Along Colorado River at river mile 243.3 R, 370 m.–113.6167301135.84593536E. Palmquist 292GCNP1 (15)21.9 (0.7)25.4 (1.4)John Wiley & Sons, Ltd. Collection site information, sample sizes, and real‐time qPCR threshold cycle (Ct) values. Site ID = code used for reference in main text. Site code = short, descriptive code for site. Location, elevation (a.s.l.) = full description of site. Locations are based on river miles from Lees Ferry and the river side when facing downstream and used by convention in all maps. For example, river mile 2.1L indicates the site is 2.1 miles downstream of Lees Ferry and on the left side of the river. Negative values indicate distance upstream of Lees Ferry. Location coordinates x = x coordinate of location in decimal degrees (WGS, 1984), y = y coordinate of location in decimal degrees (WGS, 1984). Collector = collector name and if available, specimen number. Herbarium = herbarium housing the voucher collection. Herbarium codes: GLCA = Glen Canyon National Recreation Area herbarium, GCNP = Grand Canyon National Park herbarium, DES = Desert Botanical Garden herbarium, Hualapai = Hualapai Indian Tribe herbarium. N qPCR (SSR) = sample sizes for qPCR and simple sequence repeat (microsatellite, SSR) analysis. Ct mean (SD) = Ct means and standard deviations for the AMAU and AMER probes (NA = no amplification). Leopard Frog Marsh −9.0 R Lunch Beach −7.1 L Hidden Slough −6.6 R Lees Ferry −0.1 R Genetic analysis Leaf tissue for genetic analysis was collected from 83 stands of Phragmites australis ranging from Arches National Park, Utah to the western end of Grand Canyon National Park, Arizona with the majority of samples collected in Glen Canyon National Recreation Area and Grand Canyon National Park (Figure 1, Table 1). Fifteen stands were sampled around Lake Powell and in tributaries upstream of Glen Canyon Dam. Downstream of Glen Canyon Dam, 63 stands along the Colorado River between Glen Canyon Dam and Lake Mead and in four tributaries (Paria River, Nankoweap Creek, Bright Angel Creek, Garden Creek) were tested. Five stands in Arches National Park were also tested. Samples were collected from one to 15 ramets from each stand. When more than one was collected, the samples were spread across the full area of the stand. Leaf samples were dried and stored in silica gel, and total genomic DNA was extracted from dried leaf tissue using QIAGEN DNeasy Plant MiniKits (Hilden, Germany). We tested whether samples were native Phragmites, nonnative Phragmites, a hybrid of native and nonnative Phragmites, or a third lineage, Phragmites australis subsp. berlandieri (E. Fourn.) Saltonstall & Hauber (“Gulf Coast Phragmites”). Gulf Coast Phragmites occurs along the lower Colorado River and may be expanding in the southern part of Arizona (Meyerson et al., 2010a; Lambert et al., 2016). Hybrids of native and nonnative Phragmites occur west of the study area (Saltonstall et al., 2016). We acquired leaf samples of known nonnative, native, Gulf Coast, and a hybrid of native and nonnative Phragmites from Dr. Adam Lambert (University of California, Santa Barbara, Santa Barbara, CA, USA) for positive controls (Table 1; Lambert et al., 2016). We used the protocol of Lindsay et al. (2023) for real‐time quantitative PCR (qPCR) to evaluate subspecies (native, nonnative, or Gulf Coast). One sample per site was tested, except for 11 sites where two samples per site were tested to check for mixed stands (Table 1). Samples used were from the main body of the stand. All qPCRs were run on an Applied Biosystems QuantStudio 5 Real‐Time PCR System (Thermo Fisher Scientific, Waltham, MA, USA) in 10‐µL volumes with a cover temperature of 105°C. Three technical replicates of each sample, positive controls, and negative controls were run. All three assays (AMER, AMAU, and AUBE) developed by Linsday et al. (2023) were tested. AMER and AMAU assays were run in multiplex reactions using 1X TaqMan Environmental Master Mix 2.0 (Thermo Fisher Scientific), 0.5 µM each of the AMER forward and reverse primers, 0.08 µM of the AMAU forward and reverse primers, 0.125 µM of the AMER and AMAU probes, and 1 µL of the DNA template. Tests of the AUBE assay, including a range of template dilutions and using up to 60 temperature cycles, found the best amplification of the positive controls to have threshold cycle (Ct) values greater than 40 (40.7–55.5) for the known Gulf Coast sample and no amplification for any of three known nonnative samples. Since all samples amplified for the AMAU assay (indicating no Gulf Coast samples), the AUBE assay was not pursued further. Reactions were analyzed in QuantStudio Design and Analysis Software v1.6.1 (Thermo Fisher Scientific) using default Ct thresholds. Amplification plots were evaluated for erroneous Ct scores (erratic readings not shaped like a curve), and as a result, nine Ct values were changed to “no detection” (n.d.). Microsatellite (simple sequence repeat) loci were then used to evaluate whether any hybrids of native and nonnative Phragmites were present in the collections and assess population structure among native Phragmites stands. Microsatellite data were generated for the same samples used for the qPCR analysis, and for 14 or 15 samples from each of 13 native Phragmites sites across Glen Canyon National Recreation Area and Grand Canyon National Park (Figure 1, Table 1). All microsatellite loci developed by Saltonstall (2003) and Meyerson et al. (2010b) were screened for reliable amplification, variability, and repeatable scoring. Of these, seven were amplified (PaGT4, PaGT8, PaGT9, PaGT11, PaGT13, PaGT14, PaGT21). Polymerase chain reactions were conducted in 10‐µL volumes on a 384‐well Veriti Thermal Cycler (Applied Biosystems, Foster City, CA, USA). Each reaction mixture contained 1X Phusion Hot Start II Green Master Mix (Thermo Fisher Scientific), 500 nM each of the forward (tailed including fluorescent tail oligonucleotide) and reverse primers, 1 µL of template DNA (unstandardized, neat from extraction), and molecular‐grade water. The PCR program followed the recommended sequence for Phusion Hot Start II with different annealing temperatures for different primer pairs—initial denaturation of 98°C for 30 s; 35 cycles of 98°C for 10 s, annealing temp (56–64°C) for 30 s, and 72°C for 20 s; final extension of 72°C for 10 min. Annealing temperatures for each of the primers were 56°C for PaGT9 and PaGT14; 60°C for PaGT8, PaGT11, PaGT13, PaGT21; and 64°C for PaGT4. Replicate samples, positive controls, and no‐template controls were included. Fragment analysis was conducted on an ABI 3730XL Genetic Analyzer and the associated GeneMapper Software ver. 6 (Applied Biosystems) using GeneScan LIZ500 internal size standard (Applied Biosystems). Phragmites is polypoid, but the multilocus genotypes (MLGs) of native and nonnative Phragmites can be analyzed as diploid data (e.g., Kettenring and Mock, 2012; Lambert et al., 2016; Tippery et al., 2020). The native and nonnative Phragmites ramets and the known hybrid sample did not have more than two alleles at a locus. To assess whether hybrids between native and nonnative Phragmites were present in the study area, we analyzed the microsatellite‐derived MLGs using principal coordinates analysis (PCoA) implemented in R version 4.5.1 (R Core Team, 2025) and the R package polysat with the Lynch distance matrix (cmdscale function) (Clark and Jasieniuk, 2011). All samples for which microsatellite data were available (individual samples from most stands plus 14 or 15 samples from focal population structure stands, N = 264) were included in the hybrid assessment. To determine genetic structure and differentiation across the study area, samples from the 13 native Phragmites stands with extra sampling (14 or 15 samples) were used. Four samples were missing data, so were removed from analyses for a sample size of 190 for these analyses. These microsatellite data were analyzed using k‐means clustering, discriminant analysis of principal components (DAPC), and analysis of molecular variance (AMOVA). Phragmites is a clonal plant that spreads extensively through rhizomes, so we implemented clone correction (removing identical MLGs from collection sites) and assessed linkage disequilibrium among the loci using the R packages adegenet (Jombart, 2008) and poppr (Kamvar et al., 2014) before the analyses. The clone‐corrected data (48 MLGs) had significant linkage disequilibrium, requiring statistical analyses that do not assume that loci are unlinked. Thus, k‐means clustering and DAPC were used. The function find.clusters (adegenet) was run using all principal components and k from 1 to 13 (the number of sample sites). The Bayesian information criteria (BIC) were lowest at k = 12. To evaluate admixture among the identified groups, the function dapc (adegenet) was used where all principal components and two eigenvalues were retained. To explore variability in results based on the number of groups selected, we ran DAPC analyses for k = 10, 11, and 13 clusters. An AMOVA was run to evaluate how genetic variability is partitioned between sample sites and within sample sites using the function poppr.amova (poppr), followed by a significance test (function randtest function in R package ade4; Dray and Dufour, 2007) with 9999 permutations. Weir and Cockerham F st was calculated in the R package hierfstat (function wc, Goudet, 2005), where values range from 0 to 1 and values near 1 indicate high differentiation among sites. Diversity statistics were calculated for non‐clone‐corrected data for each site and across all sites in the R package poppr (functions poppr and private_alleles; Kamvar et al., 2014)—observed and expected number of MLGs, number of private alleles, evenness, Simpson's diversity index, and expected heterozygosity. Additionally, genet richness was calculated as described by Kettenring and Mock (2012) for comparison with native Phragmites elsewhere (MLG/samples). Morphological analysis We compared morphological variation of regional native Phragmites to previous morphological assessments of nonnative Phragmites (Allred and Saltonstall, 2007; Swearington and Saltonstall, 2012; McTavish et al., 2023). We assessed qualitative and quantitative morphological characters found to reliably distinguish native and nonnative Phragmites elsewhere (Allred and Saltonstall, 2007; Swearington and Saltonstall, 2012; McTavish et al., 2023)—old stem leaf retention, leaf length, ligule base height, ligule full height, lower glume length, upper glume length, lemma length, the combined measurement of leaf length and lower glume length, stem color, and stem texture (Table 2). Based on field observations in the study area (K. Quinn, National Park Service Southeast Utah Group, personal communication), we also recorded how tightly the leaf sheath clasps the stem (tight or loose), if the community was a Phragmites monoculture or mixed with several plant species, and a previously unassessed morphological feature, axillary buds (Figure 2). Native Phragmites in the region has a small, flat flap of tissue arising from the culm node (axillary bud) that can be densely hairy (K. Quinn, National Park Service Southeast Utah Group, personal communication; Mesa County Noxious Weed & Pest Management, 2016). Axillary buds are observable on both living and past‐years’ stems. We recorded whether an axillary bud was found on any node, and if so, if the axillary bud had a tuft of hairs or not. Stem color, leaf retention, leaf snugness, and mixed or monoculture communities were recorded in the field and were based on assessments of the stand rather than an individual stem. All other characteristics were recorded from collections now housed at the herbaria of Grand Canyon National Park, Glen Canyon National Recreation Area, and the Hualapai Indian Tribe of the Hualapai Indian Reservation (Arizona). Sample sizes varied for each character based on the availability of plant parts (Table 2). All characteristics were recorded later in the summer (approximately July through October) when living stems were fully grown and stems from previous years were still available. Table 2Morphological variables assessed for regional variability in native and nonnative Phragmites australis stands confirmed with genetic data, including the abbreviation used in figures, its description, the units or categories, and sample sizes.AbbreviationVariable descriptionUnit or categoriesSample sizeLeaf retentionProportion of culm covered by leaf sheaths from previous years0.0–1.0 Native: 74 Nonnative: 4 Leaf lengthLength of a fully developed leaf blade, excluding the sheath. Average of 2 blades.cm Native: 63 Nonnative: 4 Ligule base heightHeight of the dark tissue of the ligule, excluding the hairy fringe, measured under a microscope. Average of 3 ligules.mm Native: 62 Nonnative: 2 Ligule full heightHeight of the center of the ligule, including the dark tissue and hairy fringe, measured under a microscope. Average of 5 ligules.mm Native: 62 Nonnative: 2 Lower glume lengthMean length of the lower glume measured under a microscope. Average of 5 glumes.mm Native: 57 Nonnative: 2 Upper glume lengthMean length of the upper glume measured under a microscope. Average of 5 glumes.mm Native: 57 Nonnative: 2 Lemma lengthMean length of the lemma measured under a microscope. Average of 5 lemmas.mm Native: 57 Nonnative: 2 CombGlLfCategory indicating the combined leaf length and lower glume length.1 = Lower glume <4.6 mm AND leaf length >37 cm; 2 = Lower glume <4.6 mm AND leaf length <37 cm; 3 = Lower glume >4.6 mm AND leaf length any size Native: 56 Nonnative: 2 Stem colorOrdinal scale of greenness or redness of the lower internodes of culms.1 = no red, 2 = some tinges of light red, 3 = patches of dark red covering <50% of internode, 4 = >50% of internode covered by dark red Native: 75 Nonnative: 4 Stem textureOrdinal scale indicating how smooth and shiny or dull and ridged the stem is.1 = dull, deeply ridged, 2 = lightly coarse/ridged, gentle ridges, 3 = shiny, smooth Native: 62 Nonnative: 2 CommunityCategory indicating if the stand is a monoculture or with other plant species.1 = monoculture, 2 = community with a mix of plant species Native: 74 Nonnative: 4 Leaf snugnessCategory indicating how tightly the leaf sheaths clasp the stem.1 = tight, 2 = loose Native: 74 Nonnative: 4 Node appendageOrdinal scale indicating the state of the axillary bud.1 = no axillary bud, 2 = axillary bud present, not hairy, 3 = axillary bud present with a tuft of hair. Native: 62 Nonnative: 2 John Wiley & Sons, Ltd. Morphological variables assessed for regional variability in native and nonnative Phragmites australis stands confirmed with genetic data, including the abbreviation used in figures, its description, the units or categories, and sample sizes. Native: 74 Nonnative: 4 Native: 63 Nonnative: 4 Native: 62 Nonnative: 2 Native: 62 Nonnative: 2 Native: 57 Nonnative: 2 Native: 57 Nonnative: 2 Native: 57 Nonnative: 2 Native: 56 Nonnative: 2 Native: 75 Nonnative: 4 Native: 62 Nonnative: 2 Native: 74 Nonnative: 4 Native: 74 Nonnative: 4 Native: 62 Nonnative: 2 Table 3Genetic diversity and differentiation statistics for Phragmites australis subsp. americanus collection sites derived from microsatellite loci. Site ID and Site code align with Table 1, except for average values which are given in the row labeled “Average.” Segment indicates which river, creek, or reservoir is near the sample site. N = number of ramets sampled, MLG = number of multilocus genotypes, eMLG = estimated number of multilocus genotypes per site adjusted for sample size, Au = number of private alleles, Even = Evenness, Simp = Simpson's diversity index, He = expected heterozygosity (Nei's genetic diversity), GRich = genet richness.Site IDSite codeSegment N MLGeMLGAuEvenSimp H e GRich12Escalante‐Neon CanyonEscalante River1532.9300.670.420.350.213 San Juan RM 52 San Juan River1422.0010.520.130.320.1421Warm Creek Bay SE CanyonLake Powell1511.000—00.150.0728Leopard Frog Marsh −9.0 RColorado River1532.9300.580.340.230.231 Lunch Beach −7.1 L Colorado River1565.8040.840.760.350.441Vasey's Paradise 32.2 RColorado River1444.0010.590.460.330.2945NankoweapNankoweap Creek1444.0020.590.460.310.2949Across from the LCR 62.0 RColorado River1587.8700.950.860.380.5358Bright Angel CreekBright Angel Creek1511.000—00.150.0759Havasupai GardensGarden Creek1554.8030.570.520.30.3360Granite Camp 93.8 LColorado River1422.0000.520.130.310.1468Stairway 171.6 RColorado River1422.0000.980.490.340.1483243.3 RColorado River1576.8000.880.820.30.47Average3.693.630.850.70.420.290.25John Wiley & Sons, Ltd. Genetic diversity and differentiation statistics for Phragmites australis subsp. americanus collection sites derived from microsatellite loci. Site ID and Site code align with Table 1, except for average values which are given in the row labeled “Average.” Segment indicates which river, creek, or reservoir is near the sample site. N = number of ramets sampled, MLG = number of multilocus genotypes, eMLG = estimated number of multilocus genotypes per site adjusted for sample size, Au = number of private alleles, Even = Evenness, Simp = Simpson's diversity index, He = expected heterozygosity (Nei's genetic diversity), GRich = genet richness. San Juan RM 52 Lunch Beach −7.1 L Axillary buds at culm nodes on senesced (A) Phragmites australis subsp. americanus and (B) Phragmites australis subsp. australis. Note the tufts of hairs on the axillary buds of P. australis subsp. americanus, which can be sparser or thicker, indicated by blue arrows (A). Note the lack of an axillary bud or a glabrous axillary bud on P. australis subsp. australis indicated by blue arrow (B), though the nodes can have sparse hairs even with a glabrous axillary bud. Image credit: Morgan Ford. To compare measured, continuous variables from native Phragmites to the mean and standard deviations for nonnative Phragmites presented by McTavish et al. (2023), we conducted z‐tests for each continuous variable using the R package BSDA (Arnholt and Evans, 2022). We then applied a Bonferroni correction to our target alpha of 0.05 for seven tests, resulting in a new alpha of 0.007. RESULTS Genetic analysis Of the 83 Phragmites australis sites tested in this study, four have nonnative Phragmites (Figure 1, Table 1). Two of these stands are in Arches National Park in Courthouse Wash, which also contains native Phragmites. The other two are near the confluence of the Paria River and the Colorado River along the border of Glen Canyon National Recreation Area and Grand Canyon National Park (Figure 1, Table 1). With the exception of the two stands near the mouth of the Paria River, the extensive stands occurring along the Colorado River between Glen Canyon Dam and Lake Mead are all native. Isolated populations in remote, canyon tributaries and around Lake Powell are also native. Mean Ct values and standard deviations are presented in Table 1. Analysis of microsatellite‐based MLGs from across all sample locations did not indicate that hybrids are present in the study area (Figure 3). All samples identified as native by qPCR clustered with the known native MLGs. Samples identified as nonnative by qPCR and known nonnative samples clustered separately from the native samples. The known hybrid sample was located between the two groups, as expected. Principal coordinate analysis (PCoA) of 264 Phragmites australis multilocus genotypes derived from microsatellite loci. Points represent individuals and are labeled by color: Phragmites australis subsp. australis positive control (Nonnative) = dark green, individuals from the study area identified as P. australis subsp. australis by qPCR (Nonnative Study Area) = light green, hybrid of P. australis subsp. australis and P. australis subsp. americanus positive control (Hybrid) = pink, P. australis subsp. americanus positive controls (Native) = brown, individuals from the study area identified as P. australis subsp. americanus by qPCR (Native Study Area) = orange. The k‐means cluster analysis indicated that 12 is the most likely number of genetic groups, and DAPC showed extensive admixture (gene flow) among them (Figure 4). Analyses using k = 10, 11, or 13 groups had DAPC results qualitatively equivalent to k = 12, where admixture was extensive across all groups. The AMOVA indicated that 36.0% of the total genetic variation was explained by differences among sample sites and 64.0% was explained by differences among individuals (P < 0.001). The Weir and Cockerham F st across sites was 0.17. Bar plot illustrating the posterior probability of assignment to each k‐means group based on discriminant analysis of principal components of clone‐corrected microsatellite data, where repeated multilocus genotypes within a sample site were removed, resulting in a total of 48 data points included. Samples are represented by vertical bars and are grouped by their collection sites (numbers correspond to site IDs in Tables 1, 31, 3), which are arranged from upstream to downstream from left to right. Of the 190 samples from the focal population‐genetics sample sites, 40 multilocus genotypes (MLG, unique genotypes) were identified, some of which were shared across collection sites. Stands of native Phragmites had between one and eight MLGs (Table 2) with two sites having only one clone (Table 2: sites 21, 58). The highest diversity sites were located along the Colorado River downstream of Glen Canyon Dam (Table 2: sites 31, 49, 83). Stands with unique alleles were spread across the study area, but the site that had the most private alleles (alleles only found at one collection site) is on the Colorado River near Glen Canyon Dam (Table 2: site 31). Genet richness ranged from 0.07 to 0.53 (mean = 0.25, Table 2). Nei's genetic diversity (H e, Kamvar et al., 2014) ranged from 0.15 to 0.38 (mean = 0.29, Table 2). Morphological analysis When we compared our data for continuous morphological variables for native Phragmites to published morphological ranges for nonnative Phragmites (Allred and Saltonstall, 2007; Swearington and Saltonstall, 2012; McTavish et al., 2023) and two or four nonnative samples from the study area, none of the variables perfectly distinguished native Phragmites from nonnative Phragmites (Figure 5). When measured values for native Phragmites were statistically compared to published means and standard deviations of nonnative Phragmites, all continuous variables other than the upper glume length had P‐values less than 0.001 (Figure 5). Old‐stem leaf retention was lower in native Phragmites than for nonnative Phragmites. Old‐stem leaf retention for native Phragmites in the Grand Canyon region was higher than reported for native Phragmites in northern latitudes of North America (McTavish et al., 2023) but was still generally less than that for nonnative Phragmites (Figure 5; Appendix S1: Figure S1). Similarly, the range of ligule base heights of the study area native Phragmites overlapped with that of nonnative Phragmites but was typically considerably larger (Figure 5). Thus, the proportion of culm covered by leaf sheaths from previous years (old stem leaf retention) and ligule base height can be useful traits, despite the lack of clearly distinct continuous morphological variables. Mean, SD (represented by error bars), and range (minimum and maximum values represented by points) of continuous Phragmites australis morphological traits. Measurements for native Phragmites from the study area are labeled “native” and represented by the black points and error bars with green range points. Reference means, SDs, and ranges from McTavish et al. (2023) are plotted for nonnative Phragmites (non*) and native Phragmites (native*) and illustrated in gray. Measurements for nonnative Phragmites from the study area (N = 2 or 4) are plotted in blue over the reference values. Z‐test statistics and P‐values are given for each comparison between study area native Phragmites and published means and SDs. For ordinal and categorial variables, no variables perfectly separated the two subspecies (Figure 6). All but two of the native stems had identifiable axillary buds, either hairy or not, while the two nonnative stems that could be assessed for this feature did not (Figure 6: AxillaryBud). The majority of native stems had hairy axillary buds. The majority of native stands (67) had more loosely attached leaf sheaths, whereas the nonnative stands (4) had tightly attached leaf sheaths (Figure 6: LeafSnugness). Predominantly red internodes on the lower portion of culms were restricted to native Phragmites; however, the lack of red internodes does not guarantee the stand is nonnative (Figure 6: Stemcolor). Thus, the presence of a hairy axillary bud, leaf snugness, and stem color are the most useful traits, despite the lack of clearly distinct ordinal and categorical variables. Ordinal and categorical morphological characters for individuals of native and nonnative (non) Phragmites australis from the study area. Bars represent the proportion of individuals in each category (colors). Counts within each category are shown in text. Blue indicates a category previously associated with nonnative Phragmites. Greens and orange are categories previously associated with native Phragmites. In field observations, two other features were noted as differing between native and nonnative Phragmites. First, native Phragmites bloomed later than nonnative Phragmites in the study area and had lighter‐colored seed heads (Appendix S1: Figures S2–S9). Native stands started flowering in September (sometimes starting in late August) and flower into October (Appendix S1: Figures S2–S7). Elsewhere, nonnative Phragmites flowers in August and September (Swearington and Saltonstall, 2012), but in northern Utah and other parts of the western United States, the flowering time is earlier and centers around July and August (Cronquist et al., 1977; see southwestern herbarium collections available through SEINet, 2026). In the study area, nonnative Phragmites had much darker brown seed heads that flowered starting in June and continuing through early August (Appendix S1: Figures S8, S9). Secondly, as noted elsewhere (Swearington and Saltonstall, 2012), the leaf color of the nonnative stands near the confluence of the Paria River and Colorado River was darker and more blue‐green than the lighter, yellow‐green native stands (Appendix S1: Figure S10). This leaf color difference can be noticed in the field, but is difficult to represent otherwise. DISCUSSION Native Phragmites thrives in desert wetlands and along a dam‐regulated river The greater Grand Canyon region supports mostly native Phragmites and may be one of the few remaining places in the continental United States where native Phragmites is a dominant species in wetland habitats. In contrast to invaded parts of the United States where human development and hydrological alteration have promoted nonnative Phragmites (Kettenring et al., 2012; Meyerson et al., 2025), Glen Canyon Dam operations are supporting extensive stands of native Phragmites downstream along the Colorado River in addition to stands in the desert seeps, springs, and tributaries. Higher baseflows, greater consistency of water availability, and lack of large flood events in the post‐dam Colorado River allowed native Phragmites to colonize wide reaches with low‐angle shorelines where sand and silts are deposited (Butterfield and Palmquist, 2024; Palmquist et al., 2025c). While not favored by daily hydropower tides, Phragmites tolerates fluctuating flows (Butterfield and Palmquist, 2025) as long as water and sandy substrates are available (Butterfield and Palmquist, 2024; Palmquist et al., 2025c). Phragmites does not have strong niche preferences for specific seasonal flow patterns (Butterfield et al., 2023), an advantage in a system where flows are designed for water delivery and power generation and no longer follow natural flow patterns (Palmquist, et al., 2025b). However, this is a sediment‐limited river—because the presence of the dam has reduced sediment supply to the downstream river by 95% (Topping et al., 2000, 2003)—and native Phragmites’ primary habitats, large sandbars, are eroding (Hazel et al., 2022; Palmquist et al., 2023). Despite efforts to maintain these sandbars using periodic, carefully planned floods (Hazel et al., 2022), long‐term declines in Colorado River hydrology may lead to implementation of fewer of these sand‐managing flows (Salter et al., 2025), potentially reducing native Phragmites habitat. The occurrence of nonnative Phragmites near the confluence of the Paria and Colorado Rivers as well as upstream of Lake Powell could change the relative proportions of native and nonnative Phragmites over time. Nonnative Phragmites is most likely a recent addition to the flora of the Grand Canyon region. Other regional studies document the recent spread of nonnative Phragmites into this area (Meyerson et al., 2010a; Kulmatiski et al., 2011; Lambert et al., 2016) and regular monitoring of riparian communities occurs in the area (Palmquist et al., 2018). The reason for the current dominance of native Phragmites could simply be that nonnative Phragmites only recently arrived. If nonnative Phragmites is beginning to spread and expansion patterns follow those in other parts of the continental United States, it could spread along the Colorado River between Glen Canyon Dam and Lake Mead fairly rapidly (e.g., 30 years in northern Utah, Kulmatiski et al., 2011) potentially replacing the current stands of native Phragmites. Summer flash floods from the Paria River and the periodic sand‐managing floods could promote the spread of seeds and rhizomes (Hazel et al., 2022). These floods are small and infrequent in comparison to pre‐dam floods, allowing riparian vegetation, including Phragmites, to rapidly recover and establish (Palmquist et al., 2023; Palmquist et al., 2025b). The hydrologic restriction and less‐variable hydrological conditions (as compared to pre‐dam hydrology) could facilitate the establishment and growth of nonnative Phragmites (Meyerson et al., 2025). If the habitat characteristics are suitable for nonnative Phragmites, there could be a loss of native Phragmites and rapid expansion of the nonnative in the near future. Patterns of nonnative Phragmites spread in the drylands of North America could differ from mesic regions. The flash‐flood‐prone ecosystems and comparatively small springs of the Southwest (Stevens and Meretsky, 2008) may not be ideal for the subspecies known for aggressive colonization of large, stable wetlands (Meyerson et al.,2010a, 2025; Long et al., 2017a). Nonnative Phragmites tends to spread along linear wetlands near roadsides and agricultural ditches (Maheu‐Giroux and de Blois, 2007; Jodoin et al., 2008; Kirk et al., 2011; Kettenring and Mock, 2012) and is often associated with increases in nitrogen or sediment related to human disturbance and can be somewhat tolerant of saline conditions (Kettenring et al., 2012; Lambert et al., 2016; Long et al., 2017a; Abella et al., 2024). In the drylands of North America, springs are islands separated by arid landscapes and roadsides do not support wetland habitats, so rivers and irrigation canals are the most likely conduits for water‐dependent plant species (Kettenring et al., 2012; Palmquist et al., 2021). In contrast to landscape modifications elsewhere, this regulated river is limited in sediment and nutrients, including nitrogen, and salinity has not increased (Topping et al., 2000, 2003; Deemer et al., 2023). These conditions could favor native Phragmites, limit the spread of nonnative Phragmites, and/or allow the two subspecies to coexist (Lambert et al., 2016). Indeed, at present, native and nonnative Phragmites are coexisting in Arches National Park and in the Las Vegas Wash drainage (Saltonstall et al., 2016). In Lake Powell, however, deltaic sediments at river inflows resemble typical nonnative Phragmites habitat. These deltas occur where the Colorado River and tributaries flow into the reservoir and deposit sediment. As water levels in the reservoir have lowered during recent decades, many of these fine‐grained, nutrient‐rich deltaic sediment deposits are subaerially exposed and/or periodically inundated (Deemer et al., 2025; Kasprak et al., 2026) and may provide suitable habitat for nonnative Phragmites. Hybrids between nonnative and native Phragmites are a management concern near Lake Mead (Abella et al., 2024) and could lead to the loss of pure native Phragmites through genetic swamping. Hybrids were not detected in the study area, indicating that those near Lake Mead have not spread upstream into the Grand Canyon. The conditions that lead to viable and dominant hybrids are unknown (Saltonstall et al., 2016), so it is also unknown if new hybrid lineages are likely to form in the Grand Canyon region. However, the rarity of hybrids in general and the staggered flowering periods of the two subspecies in this area (flowering starts in June for nonnative and September for native) may reduce the likelihood of hybrid spread. Genetically diverse native Phragmites can be sourced regionwide for restoration purposes The large stands of native Phragmites along the Colorado River and the smaller, isolated stands in tributaries and around Lake Powell are genetically diverse with extensive gene flow. Nei's gene diversity (H e) in the study area is similar to Phragmites in its native ranges in Australia, China, Uganda, and some parts of Europe (Guo et al., 2003; Liu et al., 2018; Qiu et al., 2021). Genet richness for native Phragmites in this study is similar to that of native Phragmites north of the study area (Kettenring and Mock, 2012), but less than reported in China (Guo et al., 2003). This reasonably high genetic diversity makes the native Phragmites stands in the greater Grand Canyon region suitable for collection and propagation for restoration purposes. Native Phragmites in the study area does not exhibit strong genetic structuring, but each stand is unique and somewhat differentiated. The F st value observed here (0.17) is similar to that of European Phragmites (0.15), but the among‐site differentiation (AMOVA, 36.0%) is considerably higher than in European populations (14.7%) (Lambertini et al., 2008). This degree of among‐site differentiation is like that of populations across all of Mexico (Colin and Eguiarte, 2016). Phragmites australis in its native range can exhibit significant site‐level genetic structuring over short distances due to environmental variability and despite high gene flow (Qiu et al., 2021). This genetic pattern is similar to that of another common, clonal riparian species along the Colorado River in the area, Salix exigua Nutt. (coyote willow) (Palmquist et al., 2021). Both species have no regional genetic structure, but have substantial site‐level differentiation and similar estimates of genetic diversity (S. exigua H e = 0.35). The same landscape‐level connectivity along a large river appears to be facilitating gene flow across broad areas, but the patchy distributions and environmental variability exhibited by both seem to be increasing site‐level differentiation (Palmquist et al., 2021; Qiu et al., 2021). Native Phragmites could be used in riparian restoration of degraded river and spring systems, where hydrology could still support its growth and persistence. For example, areas around Lake Powell that could support nonnative Phragmites could be considered for establishing native Phragmites in an attempt to prevent nonnative expansion. The high genetic diversity and high site‐level differentiation of native Phragmites is ideal for developing high‐diversity restoration materials. Individual stands have unique genetic properties, but a lack of structure across the region indicates that materials for restoration can be selected from any viable population and planted anywhere within the study area. Thus, propagating materials from many stands across the region retains the existing genetic variability, which can then be replanted in any location without disrupting regional genetic structure. Collecting and propagating Phragmites from the large populations along the Colorado River could maintain the existing genetic diversity and structure while also minimizing potential effects on smaller tributary stands. Southwestern native Phragmites is morphologically variable Currently, the most reliable means to morphologically distinguish native from nonnative Phragmites in the study area is the presence of hairy axillary buds, deep‐red internodes on the lower culms (illustrated by McTavish et al., 2023), and leaf sheaths that loosely clasp the culm (readily removed from the culm). Stands that have no axillary buds readily identifiable, entirely green culms, and have sheaths that are difficult to remove are nonnative stands. Stands that have intermediate characteristics but have axillary buds are likely native. Stands that have intermediate characteristics, but no apparent axillary buds, are likely nonnative. With questionable stands, measurements of the base of the ligule (not a field characteristic) could assist in identification, where a ligule base height greater than 0.25 mm would indicate a native stand. The axillary buds and leaf sheath “snugness” are readily evaluated year‐round on senesced stems. Evaluation of stem color and ligule bases requires mature, living stems and leaves. As noted in other studies, evaluating a combination of characters is good practice in field identification of Phragmites subspecies (Swearington and Saltonstall, 2012; McTavish et al., 2023). Flowering periods also appear to distinguish native and nonnative Phragmites. In the study area, nonnative Phragmites has dark brown seed heads that flower in June to early August (Appendix S1: Figures S8, S9). Native Phragmites in the study area have light tan seed heads and starts flowering in September (occasionally late August) and flowers into October (Appendix S1: Figures S2–S7). Several traits used for identification in eastern and northern parts of the continental United States (combined glume and leaf lengths, stem texture, upper and lower glume lengths, lemma length, and leaf length) are not reliably different in the greater Grand Canyon region due to the high variability observed within native Phragmites. The proportion of retained leaves is greater in native Phragmites in the study area than elsewhere due to the dry climate allowing leaf sheaths to be retained in higher proportions (Appendix S1: Figure S1). Additionally, unusual flow patterns related to river regulation, such as daily‐fluctuating flows to create hydropower, appear to reduce leaf sheath retention in nonnative Phragmites in the study area. Rather, the characteristic of how tightly or loosely the sheaths clasp the stem is more reliable, since this is separate from estimating the proportion of retained leaves. Also note that while red internodes are common in native Phragmites, as elsewhere, this characteristic can be lacking on immature culms. Wetland habitats in North American drylands vary considerably in their water availability, nutrient levels, salinity, flow patterns (or lack thereof), temperature, precipitation, and disturbance regimes (Stevens and Meretsky, 2008), and native Phragmites is genetically diverse. This combination of environmental variability and genetic diversity could be leading to the morphological variability observed here (Qiu et al., 2021; Meyerson et al., 2025). In environmentally stressful or variable conditions, genetic diversity in Phragmites australis is positively related to epigenetic diversity, with resulting phenotypic variability (Qiu et al., 2021; Chen et al., 2024). Phragmites australis growing in riverine and non‐riverine habitats can exhibit differences in functional traits, even where gene flow and dispersal are high (Liu et al., 2021). In the southwestern United States, native Phragmites occurs in a wide range of environmental conditions including riverine and non‐riverine habitats, low to high salinity, variable climates, and low to high disturbance. This diversity of habitats (environmental variability) could be leading to the notable phenotypic variability exhibited in this study. The western United States is characterized by large expanses of areas relatively inaccessible to people but connected ecologically through rivers. Spread of nonnative Phragmites could go undetected in understudied waterways and remote wetlands. The use of remote sensing for documenting distributions of native and nonnative Phragmites is a potential option for monitoring the spread of nonnative Phragmites (Rupasinghe and Chow‐Fraser, 2019; Durning et al., 2021; Elmer et al., 2021). In invaded portions of the United States, remote sensing is being used to map the extent of Phragmites in large wetlands (Abeysinghe et al., 2019; Chen et al., 2025) and could be used to understand the spatial distribution or spread throughout the Colorado River Basin. Durning et al. (2021) used classification of 20‐cm pixel resolution multispectral imagery acquired in 2013 to map Phragmites—not differentiated by subspecies—along 360 km of the Colorado River between Glen Canyon Dam and Lake Mead. Small stands in remote canyons could be more challenging to detect with remote‐sensing image analysis because the individual plants have tall thin stalks and relatively small canopy area. The 4.40 ha of Phragmites estimated by Durning et al. (2021) could be as small as 3.5 ha or as large as 6.2 ha based on the user's and producer's accuracies (Congalton and Green, 2019) of 79.8% and 71.1% reported by the study. However, dense stands can be detected in imagery with suitably fine pixel resolution and analysis of multispectral or hyperspectral imagery with visible and near‐infrared image reflectance bands can detect phenological characteristics such as timing of greenness, flowering, or development of brown seed heads (Rupasinghe and Chow‐Fraser, 2019; Elmer et al., 2021). In our study area, such phenology‐based remote sensing could be employed in the future to detect the changes associated with species spread or decline, to understand habitat occurrence of Phragmites by incorporating geomorphic and hydrological context with image classification and by leveraging phenological characteristics to more accurately detect Phragmites among other wetland and riparian species. Using very high spatial‐ and spectral‐resolution data and relatively large and dense stands of the native and nonnative flowering plants, it could be possible to differentially map native and nonnative Phragmites by using targeted image dates to detect the differences in flowering timing between the subspecies in our study area. CONCLUSIONS The Grand Canyon region and the Colorado River are sacred landscapes for many Native American Tribes and are also recognized by the general American public and the U.S. Government as land and water resources necessitating collaborative management (U.S. Department of the Interior, 1992; Stortz et al., 2018; Grand Canyon National Park Senior Executive Team, 2019). This region hosts a large and thriving native Phragmites population but is also at the forefront of nonnative Phragmites expansion. Management choices related to dam operations and nonnative control could alter the status and trends of this genetically and morphologically diverse hotspot for native Phragmites. The removal of nonnative Phragmites can be expensive and time‐consuming and often entails herbicide and/or destructive removal (Hazelton et al., 2014; Abella et al., 2024). The removal of species (native or nonnative) along waterways and in sacred places is both politically (Pueblo of Zuni, 2010) and ecologically complex (Robichaud et al., 2022; Bowe et al., 2024; Meyerson et al., 2025), such that nonnative Phragmites control may be controversial. However, nonnative Phragmites has been shown to substantially degrade ecosystems, through changes to nutrient cycling, sedimentation, species diversity, vegetation structural complexity, and other mechanisms (Dibble et al., 2013; Meyerson et al., 2025). Allowing nonnative Phragmites to expand in rare desert wetland and riparian habitats could lead to declines in biodiversity, increased fire hazard, and general ecosystem degradation (Abella et al., 2024; Meyerson et al., 2025). Decisions about how to manage native and nonnative Phragmites are currently being made (Abella et al., 2024), and these decisions could benefit from coordination among management agencies and stakeholders who connect to these landscapes (Armitage et al., 2009; Hoelting et al., 2025). While the U.S. National Park Service (NPS) is the largest land management agency along this section of the Colorado River, the Hualapai Indian Tribe of the Hualapai Indian Reservation, Havasupai Tribe of the Havasupai Reservation, and Navajo Nation all manage Colorado River and Lake Powell adjacent lands. Additionally, an advisory group of 25 stakeholders, including six regional Tribes and the NPS, discuss the operation of Glen Canyon Dam and provide input on the desired conditions of the Colorado River downstream of Glen Canyon Dam (Gloss et al., 2005; U.S. Department of Interior, 2016). Discussions among these diverse groups have started to help coordinate and develop best practices and management activities. The ability to scientifically inform these coordinating entities on how best to support native Phragmites is somewhat limited, since most North American studies have focused on nonnative Phragmites in habitats dissimilar from those in the drylands of North America (Saltonstall and Meyerson, 2016; Meyerson et al., 2025). Determining how likely nonnative Phragmites is to transform wetland and riparian habitats in the drylands of North America could facilitate discussion among stakeholders about the risks and benefits of Phragmites management in these unique ecosystems. Research into flow patterns and dam releases that could favor native over nonnative Phragmites is directly relevant to operations of Glen Canyon Dam and the management of the Colorado River ecosystem. Further, ecological studies evaluating competitive dominance of native and nonnative Phragmites under low and high nutrient, salinity, and disturbance conditions in arid and semiarid climates could inform our understanding of the risks to southwestern waterways. In places that have lost native Phragmites, it could be helpful to determine whether replacing nonnative Phragmites with native Phragmites, as opposed to other species, could improve restoration outcomes (Abella et al., 2024). The genetic diversity of native Phragmites presents opportunities to explore genotypes that could outcompete nonnative Phragmites and the possibility of rapid evolution and adaptation (Parmesan, 2006; Sgrò et al., 2011; Qiu et al., 2021). Knowing the conditions under which native Phragmites can outcompete nonnative, whether they be environmentally or genetically based, could provide options for riparian and wetland construction and restoration. AUTHOR CONTRIBUTIONS Conceptualization: E.P., J.S., B.B., H.F.; methodology: E.P., C.D., T.B., B.D., J.S.; data collection: E.P., C.D., T.B., B.D., formal analysis: E.P.; data curation: E.P., C.D., T.B., B.D.; writing original draft preparation: E.P.; review and editing: E.P., J.S., C.D., T.B., B.D., B.B., H.F.; visualization: E.P.; project administration: E.P., B.B.; funding acquisition: E.P., B.B., B.D. Supporting information Appendix S1: Native Phragmites australis subsp. americanus and nonnative Phragmites australis subsp. australis in the study area.